Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
308
Appendix 1 – Terms and Concepts in Model Estimation
Any quantities calculated from measurements disturbed by noise will only be estimates
of the unknown parameters of the system being modelled. The quality of a given
estimate can be assessed by its:
• Bias – the systematic error on the value obtained
• Variance - a measure of the uncertainty on the estimate.
The first- and second-order moments of a random variable x are termed its mean and
variance and defined as
}{)( xEx =µ (1A-1)
}}){{()( 22 xExEx −=σ (1A-2)
where E{•} denotes the expected value. The standard deviation is the square root of the
variance and hence has the same units as the mean value. These quantities must
normally be estimated from a finite set of N samples.
∑=
=N
iix
Nx
1
1 (1A-3)
2
1
2 )(1
1∑=
−−
=N
ii xx
Ns (1A-4)
Appendix 1 – Terms and Concepts in Model Estimation 309
which are known as the sample means and the sample variance respectively. The
variance of the sample mean is given by s2/N. The sample variance of a complex
random variable z = x + jy is given by
*
1
2 ))((1
1zzzz
Ns i
N
ii −−
−= ∑
= (1A-5)
If the variable has a complex normal distribution then the real and imaginary parts are
independent and normally distributed, with equal variances. The complex variance is
then twice the variance on either the real or imaginary parts.
For a column vector variable x, the covariance matrix is defined as
])(|)[()( TEEC xxxxx −−= (1A-6)
where the element cij of C(x) is the covariance between the elements i and j of x. It can
also be expressed in terms of a “mean of the squares” matrix and the square of the
means
TTTTTT EEEEEEEC xxxxxxxxxxxxx −=⋅+−⋅−= ][ ][)( (1A-7)
There are a number of key definitions in the classifications of estimators:
Unbiased: an estimator is unbiased if the mathematical expectation of the estimates
equals the true parameter values, p0
0)ˆ( pp =E (1A-8)
Asymptotically unbiased: if the expected value converges to the true value as the
number of measurements N increases
Appendix 1 – Terms and Concepts in Model Estimation 310
0)ˆ(lim pp =∞→
EN
(1A-9)
where N denotes the number of data points, which can be interchangeably the number of
time-domain samples or the number of frequencies.
Consistent: an estimator is consistent if Np̂ converges in some sense to p0 as N
approaches infinity. Several convergence concepts have been defined, which
convergence:
--- in mean square 0}ˆ{lim 0 =−∞→
pp NN
E (1A-10)
--- with probability one 1}ˆ{lim 0 ==∞→
pp NN
E (1A-11)
--- in probability 0 0} ˆ {lim 0 >∀=>−∞→
δδpp NN
E (1A-12)
where the probability of the bias being greater than any nonzero value δ is zero N ����
Convergence in mean square implies that the estimator is asymptotically unbiased.
Convergence in mean square and convergence with probability 1 are affinal, but not
identical, properties. Convergence in mean square and probability one both imply
convergence in probability but not vice versa. An estimator which exhibits convergence
with probability one is termed strongly consistent, while an estimator which only
exhibits convergence in probability is termed weakly consistent.
Efficient: an unbiased estimator E1 is efficient if C1<C2 for all unbiased estimators E2,
where C1 and C2 are the parameter covariance matrices, which relate the parameter
values p0 and the estimated values Np̂
} ]ˆ][ˆ[ {)ˆ( 00 pppppC −−= NNN E (1A-13)
Appendix 1 – Terms and Concepts in Model Estimation 311
The variances of the individual elements of Np̂ make up the principal diagonal of
)ˆ( NpC . A desirable property of the estimator is that the variance, or uncertainty, of the
estimates is as small as possible. However, there exists a lower bound on the covariance
matrix of an unbiased estimator, termed the Cramer-Rao bound, which is expressed as
-1 )ˆ( FpC ≥N (1A-14)
where F is the Fisher information matrix, a measure of the amount of information
present, in relation to the parameters. It is possible to construct an estimator with a
covariance matrix lower than the Cramer-Rao bound but it will be biased. The
information matrix applies to any unbiased of the parameters p using the measurements
y
00|
)|p(yln |
)|p(yln
)|p(yln E
2
2
ppppp
pp
pp
p==
∂∂−=
∂
∂
∂
∂= EFT
(1A-15)
where p(y | p) is the joint probability density function of the output in relation to the
parameter vector and the matrix of second derivatives is known as the Hessian. If p is a
d-dimensional column vector then the first derivatives will also form a d-dimensional
column vector and the Hessian will be d × d matrix.
Normal: it is highly desirable that the parameter estimates are normally distributed,
since a normally distributed variable is completely defined by its first- and second-
order moments and this allows statistical bounds to be calculated in a straightforward
manner.
Robust: if some properties of the estimator are still valid when all the a priori
assumptions are met. Most commonly this relates to a violation of the assumptions
Appendix 1 – Terms and Concepts in Model Estimation 312
regarding the noise. If the estimator can still produce a consistent estimate then it will
converge to the true value as N ����
Maximum Likelihood Estimator (MLE): the MLE is a strongly consistent and
asymptotically unbiased estimator if the true system belongs to the estimator model set.
In the presence of independent, identically distributed (iid) measurements it is also
asymptotically efficient, since it approaches the Cramer-Rao lower bound
asymptotically. If there is an estimator which reaches this lower bound it will be an
MLE. The parameter estimates are also asymptotically normal.
The primary sources for the summary presented in this Appendix were:
Åström, K. J. (1980). “Maximum likelihood and prediction error methods”, Automatica,
vol. 16, pp. 551-574.
Norton, J.P. (1986). An Introduction to Identification, Academic Press, London.
Ljung, L. (1987). System Identification - Theory for the User, Prentice Hall, Englewood
Cliffs.
Appendix 2 – Linear Gas Turbine Modelling Results 313
Appendix 2 - Linear Gas Turbine Modelling Results
________________________________________________
Table 2A-1
SMALL-SIGNAL TESTS USED FOR LINEAR GAS TURBINE MODELLING
LEVEL/TEST SIGNAL AMP
(±%Wf)
Fmin
(Hz)
Fmax
(Hz)
Bit interval
(ms)
No. bits
Level-55% NH
Test 2A
Test 3A
Test 1C
Multisine (1)
IRMLBS
GP Multisine
10
10
10
0.005
0.005
0.010
0.355
0.549
1.050
---
800
---
---
7
---
Level-65% NH
Test 4A
Test 10B
IRMLBS
IRMLBS
10
10
0.005
0.010
0.549
0.110
800
400
7
7
Level-75% NH
Test 6A
Test 7A
Test 8A
Test 11B
Multisine (1)
IRMLBS
Multisine (2)
IRMLBS
10
10
10
10
0.005
0.005
0.005
0.010
0.355
0.549
0.555
0.110
---
800
---
400
---
7
---
7
Level-80% NH
Test 12C GP Multisine 10 0.010 1.050 --- ---
Level-85% NH
Test 13A
Test 12 B
IRMLBS
IRMLBS
10
10
0.005
0.010
0.549
0.110
800
400
7
7
Level-90% NH
Test 11A
Test 12A
Test 19C
Multisine (1)
IRMLBS
GP Multisine
10
10
10
0.005
0.005
0.010
0.355
0.549
1.050
---
800
---
---
7
---
Appendix 2 – Linear Gas Turbine Modelling Results 314
0 20 40 60 80 100 120 140 160 180 200−1
−0.5
0
0.5
1
Time (s)
Am
p
(a)
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40
0.05
0.1
0.15
0.2
Frequency (Hz)
Am
p
(b)
Figure 2A-1. One period of Multisine (1) in (a) time domain and (b) frequency domain.
The time domain amplitude is normalised to one.
0 20 40 60 80 100 120 140 160 180 200−1
−0.5
0
0.5
1
Time (s)
Am
p
(a)
0 0.1 0.2 0.3 0.4 0.5 0.60
0.02
0.04
0.06
0.08
0.1
0.12
Frequency (Hz)
Am
p
(b)
Figure 2A-2. One period of Multisine (2) in (a) time domain and (b) frequency domain.
The time domain amplitude is normalised to one.
Appendix 2 – Linear Gas Turbine Modelling Results 315
0 10 20 30 40 50 60 70 80 90 100−1
−0.5
0
0.5
1
Time (s)
Am
p
(a)
0.2 0.4 0.6 0.8 1 1.20
0.05
0.1
0.15
Frequency (Hz)
Am
p
(b)
Figure 2A-3. One period of GP Multisine in (a) time domain and (b) frequency domain.
The time domain amplitude is normalised to one.
0 20 40 60 80 100 120 140 160 180 200
−1
−0.5
0
0.5
1
Time (s)
Am
p
(a)
0.5 1 1.5 2 2.50
0.05
0.1
0.15
0.2
Frequency (Hz)
Am
p
(b)
f(−3db)
Figure 2A-4. One period of IRMLBS signal in (a) time domain and (b) frequency
domain.
Appendix 2 – Linear Gas Turbine Modelling Results 316
Frequency Domain Estimation Results
Test at 55% NH
Test A3 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 482.1 54.5 0.274405 69.37 1.87 -0.2331 0.45 --- ---
1/2 69.24 53.5 0.262191 36.05 6.56 -0.8908
-0.2765
7.96
1.37
-1.0977 7.57
2/3 59.13 52.5 .026080 15.51 48.43 -0.4253
-3.4530
-0.3184
24.23
20.38
11.61
-0.5567
-4.2539
13.91
25.41
Test A3 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 476.2 54.5 0.202359 67.36 1.82 -0.2497 0.44 --- ---
1/2 97.61 53.5 0.194985 34.89 6.99 -1.0448
-0.2898
8.08
1.17
-1.1758 7.96
Test A2 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 96.1 18.5 0.26096 64.93 3.27 -0.2538 0.43 --- ---
1/2 37.09 17.5 0.257366 39.39 12.32 -0.8970
-0.2717
20.85
1.46
-0.9967 20.87
Test A2 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 110.9 18.5 0.195108 59.30 3.49 -0.2705 0.41 --- ---
1/2 55.77 17.5 0.192502 37.85 10.37 -0.8350
-0.2907
18.74
1.54
-0.9259 18.37
Appendix 2 – Linear Gas Turbine Modelling Results 317
Test C1 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 263.0 25.5 0.252527 58.77 1.414 -0.250 0.47 --- ---
1/2 26.57 24.5 0.246696 36.17 5.72 -2.31
-0.269
9.907
0.743
-2.67 10.75
2/3 24.09 23.5 0.244586 34.27 6.95 +0.195
-2.480
-0.267
70.64
10.33
0.831
+0.193
-2.89
70.94
11.39
Test C1 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 84.68 25.5 0.171629 46.09 2.01 -0.285 0.48 --- ---
1/2 50.54 24.5 0.171155 17.07 57.9 -10.4
-0.289
23.4
0.53
-15.9 48.5
2/3 43.12 23.5 0.172292 31.28 10.9 -2.29+1.9i
-2.29-1.9I
-0.284
21.9
21.9
0.73
-2.49+1.87i
-2.49-1.87i
24.9
24.9
3/4 41.32 22.5 0.170802 29.1 13.44 +0.244
-2.45+1.6i
-2.45-1.6i
-0.284
77.5
22.1
22.1
0.82
+0.242
-2.71+1.48i
-2.71+1.48i
77.8
25.9
25.9
Test at 65% NH
Test A4 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 96.28 54.5 0.170403 19.17 8.65 -0.4134 0.30 --- ---
1/2 53.05 53.5 0.179336 21.27 8.11 -0.4226
-0.0485
0.50
25.78
-0.0518 25.07
Appendix 2 – Linear Gas Turbine Modelling Results 318
Test A4 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 2147 54.5 0.134446 -23.78 2.79 -0.4840 0.14 --- ---
1/2 259.4 53.5 0.137963 7.587 13.93 -0.3675
-0.9669
1.12
2.74
-0.6961 3.51
2/3 85.46 52.5 0.14252 11.93 10.93 -0.0464
-1.3418
-0.4243
13.54
4.5
1.14
-0.0492
-1.0632
13.44
4.93
Test B10 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 90.69 54.5 0.174287 26.91 9.66 -0.394 0.539 --- ---
2/3 86.83 52.5 0.174292 27.03 10.31 -0.394
0.005+3.17i
0.005-3.17i
0.54
559.53
559.53
0.003+3.1i
0.003-3.1i
698
698
Test B10 HP Shaft (41 harmonics only BW=0.81Hz)Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 67.42 39.5 0.174265 27.47 10.37 -0.394 0.546 --- ---
2/3 58.33 37.5 0.174424 23.47 13.33 -0.393
-0.081+1.8i
-0.081-1.8i
0.58
39.47
39.47
-0.08+1.8i
-0.08-1.8i
37.7
37.7
Test B10 LP Shaft (Outlier removed)Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 423.9 54.5 0.133075 -27.06 38.01 -0.482 0.36 --- ---
1/2 115.6 53.5 0.138779 13.72 10.28 -1.26
-0.4
6.66
1.52
-1.02 7.41
2/3 64.6 52.5 0.137129 12.5 11.48 -0.400
+0.228
-1.25
1.54
36.4
6.67
+0.225
-1.01
36.6
7.46
Appendix 2 – Linear Gas Turbine Modelling Results 319
Test at 75% NH
Test A6 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 739.4 18.5 0.122747 2.953 19.11 -0.5512 0.12 --- ---
1/2 29.66 17.5 0.126552 10.42 6.39 -0.2440
-0.6159
5.11
0.78
-0.2746 5.63
Test A6 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 3139 18.5 0.091591 -84.1 1.23 -0.6949 0.19 --- ---
1/2 92.08 17.5 0.095619 -4.655 44.59 -0.4461
-1.4982
1.22
2.20
-0.8080 2.69
2/3 23.25 16.5 0.097819 4.427 63.59 -1.001
-0.5252
-1.8185
17.66
2.38
3.95
-0.1061
-1.0704
17.99
4.67
Test A8 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 196.4 54.5 0.121233 10.72 4.38 -0.5522 0.18 --- ---
1/2 75.14 53.5 0.127907 12.22 4.11 -0.1388
-0.5742
13.9
0.51
-0.1514 13.7
Test A8 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 3263 54.5 0.087608 -41.26 1.80 -0.7692 0.18 --- ---
1/2 142.7 53.5 0.095151 11.49 11.51 -0.4929
-1.8592
1.07
2.09
-1.0698 2.48
2/3 88.53 52.5 0.098083 15.13 10.14 -0.0995
-0.5552
-2.1097
17.62
1.97
3.04
-0.1067
-1.2872
17.94
3.64
Appendix 2 – Linear Gas Turbine Modelling Results 320
Test A7 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 265.2 54.5 0.122594 7.164 8.05 -0.5601 0.19 --- ---
1/2 99.87 53.5 0.127526 10.38 6.17 -0.1768
-0.5927
10.6
0.68
-0.1929 10.9
Test A7 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 4365 54.5 0.089021 -4.14 1.67 -0.7672 0.19 --- ---
1/2 272.0 53.5 0.095997 3.30 34.15 -0.4511
-1.5442
1.03
1.6
-0.8318 2.17
2/3 93.3 52.5 0.099271 9 15.45 -0.0351
-0.5113
-1.7837
2.39
1.38
2.39
-0.0372
-1.0383
13.76
3
Test B11 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 167.6 54.5 0.124708 4.64 11.59 -0.554 0.25 --- ---
1/2 72.47 53.5 0.130437 6.38 8.97 -0.241
-0.604
12.9
1.50
-0.272 13.8
2/3 70.07 52.5 0.1305 10.39 21.0 -10.6
-0.592
-0.216
29.8
1.43
14.5
-10.1
-0.240
15.3
Appendix 2 – Linear Gas Turbine Modelling Results 321
Test B11 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 2906 54.5 0.086560 -17.96 3.17 -0.858 28.27 --- ---
1/2 152.7 53.5 0.096080 9.95 7.93 -1.98
-0.497
1.94
1.39
-1.1 2.70
2/3 136.2 52.5 0.098392 11.04 7.63 -2.15
-0.567
-0.162
2.82
3.72
25.79
-1.28
-0.175
4.37
26.99
Test at 80% NH
Test C12 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 253.9 25.5 0.108624 17.24 1.67 -0.6116 0.19 --- ---
1/2 250.8 24.5 0.10867 19.11 4.41 -0.6125
+10.38
23.22
0.201
+10.6 23.97
Test C12 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 4263 25.5 0.071771 -2.597 17.07 -1.0214 0.24 --- ---
1/2 95.36 24.5 0.080049 20.658 2.96 -0.4916
-1.9945
1.39
1.35
-0.9429 2.26
Appendix 2 – Linear Gas Turbine Modelling Results 322
Test at 85% NH
Test A13 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 552.1 54.5 0.084858 -5.239 10.4 -0.7480 0.20 --- ---
1/2 87.03 53.5 0.088965 4.55 16.1 -0.4181
-0.9267
5.01
1.96
-0.528 6.56
2/3 82.25 52.5 0.089196 7.20 16.6 -2.13
-0.821
-0.336
16.48
3.06
9.18
-2.03
-0.391
16.86
10.96
Test A13 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 3555 54.5 0.060413 -49.35 1.49 -1.1817 0.25 --- ---
1/2 96.87 53.5 0.065905 4.852 28.9 -0.5575
-2.2616
1.64
1.64
-0.9428 2.50
2/3 47.34 52.5 0.067243 12.309 16.65 -0.2457
0.8003
-2.7452
13.0
5.76
4.06
-0.2787
-1.4092
14.72
6.65
Test B12 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 437.4 54.5 0.0784 -2.60 11.8 -0.8360 0.29 --- ---
1/2 106.8 53.5 0.0836 1.34 29.89 -0.5296
-1.1664
5.10
3.76
-0.773 8.30
2/3 92.67 52.5 0.085129 2.836 19.44 -2.34
-0.827
-0.254
12.8
3.26
16.8
-2.14
-0.289
13.9
18.1
Appendix 2 – Linear Gas Turbine Modelling Results 323
Test B12 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 2549 54.5 0.055121 -1506 3.30 -1.494 0.31 --- ---
1/2 93.40 53.5 0.064147 6.449 11.16 -0.5716
-2.4590
2.23
1.35
-0.9726 2.89
2/3 79.79 52.5 0.065189 8.305 10.85 -0.3452
-0.8984
-2.7307
16.9
13.5
3.54
-0.4172
-1.4061
22.1
10.7
Test B12 LP Shaft (54 harmonics)
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 2502 52.5 0.055210 -15.84 3.21 -1.49 0.31 --- ---
1/2 81.54 51.5 0.064143 6.486 11.57 -2.46
-0.572
1.38
2.23
-0.974 2.91
2/3 67.52 50.5 0.065068 8.790 11.2 -2.81
-1.00
-0.380
4.33
14.6
13.1
-1.53
-0.479
12.3
18.4
Test at 90% NH
Test A11 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 438.1 18.5 0.072370 -18.71 5.45 -0.806 0.24 --- ---
1/2 13.33 17.5 0.074789 -1.77 86.2 -0.3995
-0.9858
6.20
2.0
-0.4807 7.65
Appendix 2 – Linear Gas Turbine Modelling Results 324
Test A11 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 2227 18.5 0.052836 -98.41 1.37 -1.1767 0.30 --- ---
1/2 52.68 17.5 0.056086 -27.259 10.04 -0.4360
-1.8773
2.77
1.78
-0.5908 3.52
2/3 28.46 16.5 0.056468 2.582 306.3 -0.3070
-1.0717
-3.0830
14.3
9.74
14.3
-0.3640
-1.7276
9.51
17.15
Test A12 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 803.3 54.5 0.071250 -6.13 6.59 -0.8487 0.16 --- ---
1/2 91.63 53.5 0.075407 1.6 35.34 -0.3990
-0.9973
4.95
1.16
-0.4838 5.72
2/3 78.07 52.5 0.075627 7.693 21.74 -4.29
-0.930
-0.328
19.4
1.42
6.91
-4.03
-0.380
18.4
7.61
Test A12 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 4212 54.5 0.050133 -40.27 1.52 -1.4466 0.23 --- ---
1/2 136.7 53.5 0.055793 1.56 70.74 -0.5612
-2.3458
1.85
1.19
-0.8470 2.37
2/3 98.69 52.5 0.056970 6.712 21.58 -0.2525
-0.8214
-2.6531
12.2
5.86
2.60
-0.2843
-1.2317
13.5
6.02
Appendix 2 – Linear Gas Turbine Modelling Results 325
Test C19 HP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 297.5 25.5 0.073881 6.848 4.32 -0.8438 0.30 --- ---
1/2 40.06 24.5 0.077288 11.12 3.93 -0.5921
-1.2632
4.81
5.52
-0.9112 9.79
Test C19 LP Shaft
Order K Kmin DC gain Delay
(ms)
σ% Poles σ% Zeros σ%
0/1 3386 25.5 0.045415 -0.73 49.64 -1.7264 0.25 --- ---
1/2 49.91 24.5 0.052454 17.173 3.05 -0.5974
-2.6306
2.06
0.99
-0.9478 2.50
2/3 26.27 23.5 0.054264 18.08 3.14 -0.1746
-0.7471
-2.7639
28.1
6.39
1.68
-0.1921
-1.1617
29.55
5.74
Time Domain Estimation Results
Test at 55% NH
Test A2 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.3140 --- --- 0.9878 0.0019
1/2 0.3173 0.9119 0.0062 0.9867
0.9232
0.0013
0.0029
2/3 0.2861 0.9842
0.8563
0.0071
0.0161
0.9859+0.0028i
0.9859-0.0028i
0.8738
0+0.0001i
0.3012
0.0001
Appendix 2 – Linear Gas Turbine Modelling Results 326
Test A2 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 0.7115 --- --- 0.9869 0.0019
1/2 0.2023 0.9056 0.0068 0.9859
0.9174
0.0013
0.0031
2/3 0.2031 0.9892
0.6572
0.0084
0.0496
0.9879+0.0019i
0.9879-0.0019i
0.7887
0.0+0.008i
-23-0.04i
0.0092
Test C1 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.2176 --- --- 0.9913 0.0078
1/2 0.0361 0.7304 0.0517 0.9897
0.7969
0.0027
0.0044
2/3 0.0070 0.9538
-0.5040
0.0806
-5.6035
0.9891
0.9576
0.5860
0.0430
0.1324
0.1281
Test C1 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 0.2420 --- --- 0.9897 0.0069
1/2 0.0085 0.6063 0.1733 0.9884
0.6988
0.0019
0.0863
Appendix 2 – Linear Gas Turbine Modelling Results 327
Test A3 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 3.9872 --- --- 0.9897 0.0041
1/2 0.2085 0.9144 0.0064 0.9872
0.9305
0.0026
0.0036
2/3 0.1226 0.9667
0.4594
0.042
0.1622
0.9857
0.9728
0.7120
0.0197
0.0618
0.0498
Test A3 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 2.1158 --- --- 0.9889 0.0043
1/2 0.1521 0.9014 0.0105 0.9865
0.9195
0.0034
0.0049
2/3 0.0932 0.9714
0.3877
0.0242
0.3704
0.9835
0.9782
0.7024
0.0308
0.0511
0.0278
Test at 65% NH
Test A4 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 2.8148 --- --- 0.9815 0.0057
1/2 0.4257 0.1181 3.0598 0.9797
0.5156
0.0024
0.0300
Appendix 2 – Linear Gas Turbine Modelling Results 328
Test A4 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 3.9111 --- --- 0.9735 0.0064
1/2 0.3331 0.9249 0.0042 0.9786
0.9057
0.0031
0.0072
2/3 0.1686 0.9724
0.8501
0.0525
0.0465
0.9826
0.9624
0.8288
0.0159
0.0439
0.0684
Test B10 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 0.3502 --- --- 0.9814 0.0144
1/2 0.3236 0.7713 0.4053 0.9801
0.7929
0.0218
0.23989
Test B10 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 0.7775 --- --- 0.9741 0.0126
1/2 0.3637 0.8761 0.223
0
0.9771
0.9603
0.0239
0.2129
Appendix 2 – Linear Gas Turbine Modelling Results 329
Test at 75% NH
Test A6 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.2424 --- --- 0.9735 0.0025
1/2 0.4895 0.9487 0.0057 0.9754
0.9408
0.0033
0.0028
2/3 0.1930 0.9828
0.0929
0.0169
0.7309
0.9851
0.9685
0.4566
0.0139
0.0056
0.0218
Test A6 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 9.6100 --- --- 0.9605 0.0132
1/2 0.3983 0.9397 0.0086 0.9744
0.8931
0.0032
0.0045
2/3 0.1379 0.9830
0.9095
0.0051
0.0115
0.9861
0.9636
0.8621
0.0058
0.0056
0.0046
Appendix 2 – Linear Gas Turbine Modelling Results 330
Test A8 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.0289 --- --- 0.9720 0.0030
1/2 0.6499 0.9407 0.0079 0.9753
0.9334
0.0105
0.0181
2/3 0.2064 0.9804
-0.1683
0.1567
-1.5400
0.9829
0.9690
0.3955
0.1109
0.0576
0.2307
Test A8 LP Shaft
Order Cost
Fn.
Zeros σz
(%)
Poles σp
(%)
0/1 4.9941 --- --- 0.9600 0.0092
1/2 0.3877 0.9276 0.0067 0.9723
0.8823
.00067
0.0167
2/3 0.2473 0.9438
-0.4369
0.0328
-3.0081
0.9747
0.9057
0.3215
0.0059
0.0479
0.1612
3/4 0.2359 0.96+0.1
7i
0.96-
0.17i
0.9404
0+0i
-18-3i
0.0243
0.9596+0.1725i
0.9596-0.1725i
0.9744
0.8968
0+0i
0.7+0.1i
0.0099
0.0626
Appendix 2 – Linear Gas Turbine Modelling Results 331
Test A7 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 2.2056 --- --- 0.9719 0.0031
1/2 1.4059 0.9158 0.0099 0.9740
0.9084
0.0042
0.0092
2/3 0.2566 -1.0898
0.9911
-0.6909
0.0082
0.9919
0.9711
0.1751
0.0076
0.0034
0.0236
3/4
Test A7 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 8.2479 --- --- 0.9604 0.0088
1/2 0.6078 0.9310 0.0073 0.9728
0.8863
0.0054
0.0066
2/3 0.4128 0.9806
0.8912
0.0128
0.0150
0.9850
0.9582
0.8501
0.0118
0.0145
0.0159
Appendix 2 – Linear Gas Turbine Modelling Results 332
Test B11 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.6142 --- --- 0.9758 0.0157
1/2 0.2278 -231.29 -86.32 0.9432
-0.6566
0.0035
-0.7691
2/3 0.1809 -1.1885
0.8780
-6.0024
0.8292
0.9737
0.8733
0.1134
0.0369
0.7699
2.1677
Test B11 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 6.0114 --- --- 0.9509 0.0320
1/2 0.9448 0.8844 0.0884 0.9667
0.8289
0.0286
0.0563
2/3 0.1454 -1.3392
0.9361
-1.6577
0.4007
0.9734
0.8946
0.1491
0.1162
0.3605
1.1291
Appendix 2 – Linear Gas Turbine Modelling Results 333
Test at 80% NH
Test C12 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.8760 --- --- 0.9612 0.0046
1/2 0.8436 0.8794 0.0079 0.9631
0.8524
0.0047
0.0119
2/3 0.2587 0.9745
0.4838
0.0038
0.1100
0.9790
0.9527
0.3010
0.0059
0.0076
0.0523
Test C12 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 9.2170 --- --- 0.9303 0.0171
1/2 0.8549 0.9373 0.0056 0.9641
0.8553
0.0068
0.0166
2/3 0.2974 0.9773
0.8710
0.0330
0.0262
0.9821
0.9312
0.8012
0.0252
0.0211
0.0550
Appendix 2 – Linear Gas Turbine Modelling Results 334
Test at 85% NH
Test A13 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 1.7490 --- --- 0.9657 0.0044
1/2 0.7866 0.8793 0.0139 0.9669
0.8555
0.0046
0.0051
2/3 0.1749 0.9815
0.6188
0.0294
0.0660
0.9842
0.9596
0.5253
0.0233
0.0100
0.0457
Test 13 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 9.4711 --- --- 0.9403 0.0159
1/2 0.6299 0.9348 0.0080 0.9669
0.8614
0.0063
0.0068
2/3 0.1602 0.9777
0.8862
0.0330
0.0225
0.9829
0.9426
0.8203
0.0228
0.0211
0.0434
Appendix 2 – Linear Gas Turbine Modelling Results 335
Test B12 HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 2.7365 --- --- 0.9561 0.0202
1/2 0.7123 0.6248 0.1791 0.9612
0.5006
0.0091
0.0695
2/3 0.3326 -2.4395
0.9289
-0.8562
0.1076
0.9655
0.9171
-0.2241
0.0213
0.0916
-0.1343
Test B12 LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 4.2329 --- --- 0.9320 0.0319
1/2 0.9442 0.8930 0.1800 0.9547
0.8047
0.0444
0.2220
2/3 0.1663 0.9500
0.3364
0.1598
1.6262
0.9700
0.8842
0.0682
0.1004
0.0948
5.2262
Appendix 2 – Linear Gas Turbine Modelling Results 336
Test at 90% NH
Test 11A HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 2.6982 --- --- 0.9597 0.0056
1/2 0.5422 0.9006 0.0785 0.9663
0.8349
0.0056
0.1571
1/3 0.0947 0.9818
0.7250
0.0125
0.0787
0.9838
0.9569
-0.4427
0.0098
0.0067
-1.8034
Test 11A LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 10.012 --- --- 0.9334 0.0260
1/2 0.5892 0.9457 0.0352 0.9691
0.8205
0.0123
0.1135
1/3 0.0785 0.9788
0.7982
0.0098
0.0589
0.9829
0.9314
-0.9723
0.0066
0.0160
-0.8001
Appendix 2 – Linear Gas Turbine Modelling Results 337
Test 12A HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 3.5244 --- --- 0.9590 0.0060
1/2 0.8959 0.8002 0.1078 0.9633
0.6216
0.0042
0.3540
1/3 0.2401 0.9843
0.697
0.0091
0.0790
0.9859
0.9579
-0.9905
0.0077
0.0057
-0.3915
Test 12A LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 14.6488 --- --- 0.9261 0.0284
1/2 1.2841 0.9258 0.0491 0.9632
0.7831
0.0145
0.1452
Test 19C HP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 0.6431 --- --- 0.9679 0.0056
1/2 0.2768 0.5602 1.1787 0.9693
-0.4733
0.0044
13.701
Test 19C LP Shaft
Order Cost Fn. Zeros σz
(%)
Poles σp
(%)
0/1 3.0872 --- --- 0.9357 0.0303
1/2 0.6668 0.9137 0.1279 0.9639
0.7936
0.0278
0.2822
Appendix 3 – Nonlinear Gas Turbine Estimation Results 338
Appendix 3 – Nonlinear Gas Turbine Estimation Results
________________________________________________
Table 3A-1
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 55% NH, nu=ny=2 l=2, HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 3.9105 0.5459 0.5455 0.5468 0.5460
3 76.712 1.7364 1.7364 1.7376 1.7364
4 0.1273 -0.8224 -0.8232 -0.8207 -0.8223
5 0.0515 -1.1836 -1.1846 -1.1815 -1.1835
6 0.0243 -1.4827 -1.4839 -1.4801 -1.4825
7 0.0341 -1.3473 -1.3487 -1.3443 -1.3471
8 0.0174 -1.6146 -1.6162 -1.6112 -1.6144
9 0.0180 -1.6008 -1.6026 -1.5970 -1.6006
10 0.0203 -1.5537 -1.5557 -1.5494 -1.5535
11 0.0204 -1.5507 -1.5529 -1.5460 -1.5505
12 0.0215 -1.5292 -1.5316 -1.5241 -1.5290
13 0.0202 -1.5547 -1.5573 -1.5491 -1.5544
14 0.0296 -1.4011 -1.4039 -1.3951 -1.4008
15 0.0298 -1.3974 -1.4004 -1.3909 -1.3970
Appendix 3 – Nonlinear Gas Turbine Estimation Results 339
Table 3A-2
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 65% NH, nu=ny=2 l=2, HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 9.9842 0.9207 0.9234 0.9216 0.9207
3 37.790 1.4533 1.4527 1.4546 1.4534
4 0.0270 -1.4412 -1.4420 -1.4394 -1.4411
5 0.0170 -1.6248 -1.6258 -1.6226 -1.6246
6 0.0311 -1.3850 -1.3862 -1.3824 -1.3848
7 0.0601 -1.1209 -1.1223 -1.1179 -1.1207
8 0.0272 -1.4376 -1.4392 -1.4342 -1.4374
9 0.0281 -1.4233 -1.4251 -1.4194 -1.4231
10 0.0293 -1.4072 -1.4092 -1.4029 -1.4070
11 0.0284 -1.4181 -1.4203 -1.4134 -1.4179
12 0.0300 -1.3960 -1.3984 -1.3908 -1.3957
13 0.0286 -1.4157 -1.4183 -1.4101 -1.4154
14 0.0496 -1.1951 -1.1979 -1.1891 -1.1947
15 0.0497 -1.1933 -1.1963 -1.1869 -1.1930
Appendix 3 – Nonlinear Gas Turbine Estimation Results 340
Table 3A-3
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 75% NH, nu=ny=2 l=2, HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 15.610 1.0994 1.0990 1.1003 1.0995
3 1.6026 0.1897 0.1891 0.1910 0.1898
4 0.0158 -1.6560 -1.6568 -1.6543 -1.6559
5 0.0122 -1.7574 -1.7584 -1.7552 -1.7573
6 0.0087 -1.8917 -1.8929 -1.8891 -1.8916
7 0.0158 -1.6554 -1.6568 -1.6524 -1.6552
8 0.0181 -1.6015 -1.6031 -1.5981 -1.6013
9 0.0180 -1.6022 -1.6040 -1.5983 -1.6020
10 0.0153 -1.6645 -1.6665 -1.6603 -1.6643
11 0.0156 -1.6585 -1.6607 -1.6538 -1.6582
12 0.0154 -1.6615 -1.6639 -1.6564 -1.6613
13 0.0159 -1.6481 -1.6507 -1.6425 -1.6478
14 0.0162 -1.6414 -1.6442 -1.6354 -1.6411
15 0.0159 -1.6480 -1.6510 -1.6415 -1.6476
Appendix 3 – Nonlinear Gas Turbine Estimation Results 341
Table 3A-4
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 85% NH, nu=ny=2 l=2, HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 19.553 1.1894 1.1890 1.1903 1.1895
3 127.40 1.9392 1.9386 1.9405 1.9392
4 0.0156 -1.6602 -1.6610 -1.6585 -1.6601
5 0.0096 -1.8529 -1.8539 -1.8508 -1.8528
6 0.0124 -1.7515 -1.7527 -1.7490 -1.7514
7 0.0187 -1.5878 -1.5892 -1.5848 -1.5876
8 0.0168 -1.6300 -1.6316 -1.6265 -1.6298
9 0.0177 -1.6084 -1.6102 -1.6045 -1.6082
10 0.0168 -1.6282 -1.6302 -1.6239 -1.6280
11 0.0163 -1.6403 -1.6425 -1.6356 -1.6400
12 0.0165 -1.6341 -1.6365 -1.6290 -1.6339
13 0.0164 -1.6364 -1.6390 -1.6308 -1.6361
14 0.0177 -1.6072 -1.6100 -1.6012 -1.6069
15 0.0177 -1.6058 -1.6088 -1.5994 -1.6055
Appendix 3 – Nonlinear Gas Turbine Estimation Results 342
Table 3A-5
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 55% NH, nu=ny=2 l=2 (STANDARDISED DATA) , HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 77.018 1.7380 1.7376 1.7388 1.7380
3 76.712 1.7364 1.7358 1.7376 1.7364
4 1.1231 0.0480 0.0472 0.0497 0.0481
5 0.0303 -1.3954 -1.3964 -1.3932 -1.3952
6 0.0341 -1.3482 -1.3494 -1.3457 -1.3481
7 0.0183 -1.5949 -1.5963 -1.5919 -1.5947
8 0.0188 -1.5847 -1.5863 -1.5813 -1.5846
9 0.0188 -1.5847 -1.5865 -1.5808 -1.5845
10 0.0191 -1.5765 -1.5785 -1.5722 -1.5763
11 0.0211 -1.5368 -1.5390 -1.5320 -1.5365
12 0.0199 -1.5607 -1.5631 -1.5556 -1.5605
13 0.0203 -1.5522 -1.5548 -1.5466 -1.5519
14 0.0296 -1.4011 -1.4039 -1.3951 -1.4008
15 0.0298 -1.3974 -1.4004 -1.3909 -1.3909
Appendix 3 – Nonlinear Gas Turbine Estimation Results 343
Table 3A-6
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 65% NH, nu=ny=2 l=2 (STANDARDISED DATA) , HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 38.100 1.4565 1.4561 1.4574 1.4566
3 37.790 1.4533 1.4527 1.4546 1.4534
4 0.2628 -0.5326 -0.5334 -0.5308 -0.5325
5 0.0750 -1.0331 -1.0341 -1.0310 -1.0330
6 0.0630 -1.1028 -1.1040 -1.1002 -1.1027
7 0.0403 -1.2809 -1.2823 -1.2779 -1.2807
8 0.0412 -1.2718 -1.2734 -1.2683 -1.2716
9 0.0259 -1.4564 -1.4582 -1.4526 -1.4562
10 0.0276 -1.4311 -1.4331 -1.4268 -1.4309
11 0.0306 -1.3894 -1.3916 -1.3847 -1.3891
12 0.0291 -1.4091 -1.4115 -1.4040 -1.4088
13 0.0322 -1.3683 -1.3709 -1.3627 -1.3680
14 0.0496 -1.1951 -1.1979 -1.1891 -1.1947
15 0.0497 -1.1933 -1.1963 -1.1869 -1.1930
Appendix 3 – Nonlinear Gas Turbine Estimation Results 344
Table 3A-7
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 75% NH, nu=ny=2 l=2 (STANDARDISED DATA), HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 1.5547 0.1772 0.1768 0.1781 0.1773
3 1.6026 0.1897 0.1891 0.1910 0.1898
4 0.4462 -0.3210 -0.3218 -0.3192 -0.3209
5 0.0151 -1.6744 -1.6754 -1.6722 -1.6743
6 0.0165 -1.6373 -1.6385 -1.6347 -1.6372
7 0.0244 -1.4804 -1.4818 -1.4774 -1.4803
8 0.0242 -1.4832 -1.4848 -1.4797 -1.4830
9 0.0159 -1.6500 -1.6518 -1.6461 -1.6497
10 0.0161 -1.6465 -1.6485 -1.6422 -1.6462
11 0.0152 -1.6670 -1.6692 -1.6623 -1.6667
12 0.0157 -1.6537 -1.6561 -1.6485 -1.6534
13 0.0154 -1.6621 -1.6647 -1.6565 -1.6618
14 0.0162 -1.6414 -1.6442 -1.6354 -1.6411
15 0.0159 -1.6480 -1.6510 -1.6415 -1.6476
Appendix 3 – Nonlinear Gas Turbine Estimation Results 345
Table 3A-8
STATISTICAL CRITERIA COMPUTED FOR A VALIDATION DATA
SET AT 85% NH, nu=ny=2 l=2 (STANDARDISED DATA), HP Shaft
Model size NV AIC
×104
FPE
×104
BIC
×104
LILC
×104
1 *** *** *** *** ***
2 126.47 1.9363 1.9359 1.9372 1.9364
3 127.40 1.9392 1.9386 1.9405 1.9392
4 0.2970 -0.4836 -0.4844 -0.4819 -0.4835
5 0.0218 -1.5266 -1.5276 -1.5244 -1.5265
6 0.0191 -1.5782 -1.5794 -1.5756 -1.5781
7 0.0203 -1.5543 -1.5557 -1.5513 -1.5541
8 0.0213 -1.5346 -1.5362 -1.5312 -1.5344
9 0.0165 -1.6351 -1.6369 -1.6312 -1.6349
10 0.0167 -1.6313 -1.6333 -1.6270 -1.6311
11 0.0170 -1.6230 -1.6252 -1.6183 -1.6228
12 0.0169 -1.6253 -1.6277 -1.6202 -1.6251
13 0.0179 -1.6012 -1.6038 -1.5956 -1.6009
14 0.0177 -1.6072 -1.6100 -1.6012 -1.6069
15 0.0177 -1.6058 -1.6088 -1.5994 -1.6055
Appendix 3 – Nonlinear Gas Turbine Estimation Results 346
40 60 8050
54
58
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
62
64
66
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8072
74
76
78
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
82
85
88
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18056
64
72
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 20065
75
85
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-1. Outputs of validation data sets, HP shaft. Measured output (solid), Model
1 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, , (d) IRMLBS test at 85% NH, , (e) three-level periodic
test at 58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
40 60 8050
54
58
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
62
64
66
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8072
74
76
78
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
82
85
88
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18056
64
72
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 20065
75
85
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-2. Outputs of validation data sets, HP shaft. Measured output (solid), Model
2 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, , (d) IRMLBS test at 85% NH, , (e) three-level periodic
test at 58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
Appendix 3 – Nonlinear Gas Turbine Estimation Results 347
40 60 8050
54
58
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
62
64
66
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8072
74
76
78
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
82
85
88
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18056
64
72
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 20065
75
85
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-3. Outputs of validation data sets, HP shaft. Measured output (solid), Model
3 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, , (d) IRMLBS test at 85% NH, , (e) three-level periodic
test at 58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
40 60 80
25
26
27
28
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
32
34
36
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8040
43
46
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
48
50
52
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18028
34
40
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 200
35
40
45
50
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-4. Outputs of validation data sets, LP shaft. Measured output (solid), Model
1 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, (d) IRMLBS test at 85% NH, (e) three-level periodic test at
58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
Appendix 3 – Nonlinear Gas Turbine Estimation Results 348
40 60 80
25
26
27
28
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
32
34
36
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8040
43
46
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
48
50
52
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18028
34
40
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 200
35
40
45
50
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-5. Outputs of validation data sets, LP shaft. Measured output (solid), Model
2 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, (d) IRMLBS test at 85% NH, (e) three-level periodic test at
58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
40 60 80
25
26
27
28
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
32
34
36
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8040
43
46
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
48
50
52
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18028
34
40
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 200
35
40
45
50
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-6. Outputs of validation data sets, LP shaft. Measured output (solid), Model
3 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, (d) IRMLBS test at 85% NH, (e) three-level periodic test at
58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.
Appendix 3 – Nonlinear Gas Turbine Estimation Results 349
40 60 80
25
26
27
28
Time (s)
Sha
ft S
peed
(%
NH
)
(a)
40 60 80
32
34
36
Time (s)
Sha
ft S
peed
(%
NH
)
(b)
40 60 8040
43
46
Time (s)
Sha
ft S
peed
(%
NH
)
(c)
40 60 80
48
50
52
Time (s)
Sha
ft S
peed
(%
NH
)
(d)
100 120 140 18028
34
40
Time (s)
Sha
ft S
peed
(%
NH
)
(e)
80 120 160 200
35
40
45
50
Time (s)
Sha
ft S
peed
(%
NH
)
(f)
Figure 3A-7. Outputs of validation data sets, LP shaft. Measured output (solid), Model
4 output (dashed). (a) multisine (1) test at 55% NH, (b) IRMLBS test at 65% NH, (c)
multisine (1) test at 75% NH, (d) IRMLBS test at 85% NH, (e) three-level periodic test
at 58-70% NH, (f) triangular wave + IRMLBS test at 65-85% NH.